Revisiting the small motion assumption